Skidmore College
Creative Matter
Geosciences Faculty Scholarship
Geosciences
2012
Rate and Apparent Quantum Yield of
Photodissolution
Meg Estapa
Skidmore College
Lawrence M. Mayer
Emmanuel Boss
Follow this and additional works at: http://creativematter.skidmore.edu/geosci_fac_schol
Part of the Oceanography and Atmospheric Sciences and Meteorology Commons
Recommended Citation
Estapa, M.L., Mayer L.M., Boss E., 2012. Rate and apparent quantum yield of photodissolution. Limnology and Oceanography, 57(6):
1743-1756.
This Article is brought to you for free and open access by the Geosciences at Creative Matter. It has been accepted for inclusion in Geosciences Faculty
Scholarship by an authorized administrator of Creative Matter. For more information, please contact [email protected].
Limnol. Oceanogr., 57(6), 2012, 1743–1756
2012, by the Association for the Sciences of Limnology and Oceanography, Inc.
doi:10.4319/lo.2012.57.6.1743
E
Rate and apparent quantum yield of photodissolution of sedimentary organic matter
Margaret L. Estapa,a,b,1,* Lawrence M. Mayer,a and Emmanuel Boss b
a School
b School
of Marine Sciences, University of Maine, Darling Marine Center, Walpole, Maine
of Marine Sciences, University of Maine, Orono, Maine
Abstract
We quantified rates of photochemical dissolution (photodissolution) of organic carbon in coastal Louisiana
suspended sediments, conducting experiments under well-defined conditions of irradiance and temperature.
Optical properties of the suspended sediments were characterized and used in a radiative transfer model to
compute irradiances within turbid suspensions. Photodissolution rate increased with temperature (T), with
activation energy of 32 6 7 kJ mol21, which implicates indirect (non-photochemical) steps in the net reaction. In
most samples, dissolved organic carbon (DOC) concentration increased approximately linearly with time over the
first 4 h of irradiation under broadband simulated sunlight, after higher rates in the initial hour of irradiation.
Four-hour rates ranged from 2.3 mmol DOC m23 s21 to 3.2 mmol DOC m23 s21, but showed no relation to
sample origin within the study area, organic carbon or reducible iron content, or mass-specific absorption
coefficient. First-hour rates were higher—from 3.5 mmol DOC m23 s21 to 7.8 mmol DOC m23 s21—and
correlated well with sediment reducible iron (itself often associated with organic matter). The spectral apparent
quantum yield (AQY) for photodissolution was computed by fitting DOC photoproduction rates under different
spectral irradiance distributions to corresponding rates of light absorption by particles. The photodissolution
AQY magnitude is similar to most published dissolved-phase AQY spectra for dissolved inorganic carbon
photoproduction, which suggests that in turbid coastal waters where particles dominate light absorption, DOC
photoproduction from particles exceeds photooxidation of DOC.
Recent laboratory studies have demonstrated that under
solar irradiance, significant amounts of particulate organic
carbon (POC) in suspended marine and aquatic particles
can be photochemically released as dissolved organic
carbon (DOC). Under simulated sunlight, Mayer et al.
(2006), Kieber et al. (2006), and Riggsbee et al. (2008)
induced photochemical POC dissolution (photodissolution)
in resuspended sediments in coastal and freshwater
environments. Photodissolution of POC from Louisiana
sediments leads to O2 consumption, dissolved inorganic
carbon (DIC) production, peroxide production (Estapa
and Mayer 2010), and loss of lignin, bromine, and D14Cmodern carbon from particles (Mayer et al. 2009a).
Particle-hosted photochemical reactions are not limited to
mineral-rich coastal sediments; irradiation of algal cellular
detritus results in similar photodissolution extents (Mayer
et al. 2009b). Irradiation of particles from both coastal and
oliogotrophic regimes results in efficient production of
carbon monoxide (Xie and Zafiriou 2009).
Such photochemical transformations could affect fluxes
of carbon through marine systems in a variety of ways. For
instance, burial of POC in shallow sediments is an
important removal term in coastal carbon budgets that
could be reduced by sediment resuspension and subsequent
POC photodissolution. In the overlying coastal water
column, photochemical production of DOC from suspended particulate matter (SPM) and episodically resuspended
sediments may provide DOC to microbes and to cross-shelf
* Corresponding author: [email protected]
1 Present
address: Marine Chemistry and Geochemistry Department, Woods Hole Oceanographic Institution, Woods Hole,
Massachusetts
transport (Mayer et al. 2011). Offshore, sunlight contributes to the photochemical alteration of detrital particles, as
evidenced by photochemical degradation products in lipids
from sinking and suspended particulates (Rontani et al.
2011). Photochemical reactions may also affect the cycling
of iron–organic complexes (Waite 2005).
Solar irradiance varies substantially as a function of
wavelength, time, location, and depth within the water
column, and particulate absorption coefficients in the
environment depend on highly variable suspended particle
concentrations. Quantitative estimates of particle-hosted,
photochemical reaction rates thus require knowledge of
their spectral efficiency, or apparent quantum yield (AQY).
To our knowledge, only one marine particulate photochemistry study to date reports particulate absorption
coefficients (Xie and Zafiriou 2009), and Mayer et al.
(2006) only qualitatively accounted for absorbed irradiance
within experimental suspensions. Here, we use an experimental approach to estimate the spectral AQY of one
particulate photoreaction, POC photodissolution, for a
collection of coastal SPM samples from Louisiana. We also
quantify the positive temperature-dependence of the
photodissolution rate indicated by Mayer et al. (2006).
Finally, we examine the dependence of photodissolution
rate on the particles’ light absorption coefficients, reducible
iron, and organic carbon contents.
Methods
To account for irradiance absorbed by particles, we
require a radiative transfer model that is valid for turbid
water and one that describes scalar irradiance attenuation
in terms of measureable radiometric quantities and particle
1743
1744
Estapa et al.
optical properties. Below, we derive this model from prior
photochemical work in non-scattering water. Next, we
describe our experimental design, and finally our sample
analytical techniques.
of change of net (i.e., vector) irradiance, defined as the
downward minus the upward irradiance (Enet ; Ed 2 Eu),
equals the scalar irradiance times the absorption coefficient
within the water:
Definitions—The AQY is the photochemical reaction rate
normalized to the light absorption rate of the reacting
substance (Eq. 1). The adjective ‘apparent’ is necessary in the
case of organic matter in the environment because neither
the chemical concentrations of the actual reacting molecules,
nor their relative contributions to the measured total
absorption coefficient, are known. We define the AQY as
dEnet (z)
~{at (z)|Eo (z)
dz
d½DOC
AQY(l)~ dt
Q(l)
SA
Q(l)~
V
ð0
Eo (l,z)|ax (l)dz
ð2Þ
{?
and V (m3) are, respectively, the area of the
Above, SA
upper surface and the volume of the reaction volume, Eo(l,z)
is scalar irradiance (mE m22 s21) as a function of wavelength
and depth z (m), and ax(l) is the absorption coefficient of
the reacting substance (for simplicity, we assume that all
inherent optical properties are constant with respect to
depth). If light scattering is negligible (i.e., in the absence of
particles), Eq. 2 can be integrated with respect to depth and
Q(l) computed according to Eq. 3 (following Hu et al. 2002;
wavelength notation dropped for brevity):
SA ax
|
V
at
0ð{
Enet (0 ){Enet (H)~
ð1Þ
(m2)
Q~Ed (0{ )|(1{e{at H )|
Integrating Eq. 4 over depth from just below the surface
(02) to the bottom of the finite volume H, light absorbed
within the volume (Eq. 2) is equivalent to the difference
between Enet(02) and Enet(H):
{
where d [DOC]/dt refers to the production rate of DOC
(mmol DOC m23 s21), and Q(l) (mE m23 s21) refers to the
rate of light absorption by the photoreacting substance as a
function of wavelength (l). Q(l) is the depth-integrated
product of scalar irradiance and the absorption coefficient
of the photoreacting substance, normalized to the reaction
volume:
ð3Þ
where at (m21) is the total absorption coefficient (including
water and dissolved substances), Ed (02) (mE m22 s21) is the
area-normalized rate of downward light flux to a depth just
below the surface, and H (m) is the depth of the reaction
volume. The dimensionless term (1 2 e2at H) is the fraction of
light not absorbed above depth H.
In the case of water with a non-negligible particle
concentration, light is both scattered and absorbed, and the
computation of Q must also account for scattering losses of
light from the experimental volume. Figure 1 shows a
schematic of the experimental volume and the irradiance
quantities discussed in our derivation of Q for scattering
water, below. In a horizontally infinite system, or one
bounded by specularly reflecting (mirror-like) walls, these
depth-integrated scattering losses equal the sum of light lost
from the upper air–water interface and absorbed by the
bottom. Gershun’s law, derived for a plane-parallel
medium (Eq. 4; Gershun 1939), states that the depth rate
ð4Þ
Eo (z)|at dz
ð5Þ
H
Conservation of energy requires that net irradiance remain
constant above and below the air–water interface of the
experimental volume (Enet(02) 5 Enet(0+); Fig. 1), so after
defining above-water reflectance as the ratio of upward to
downward planar irradiance (R(0+) 5 Eu(0+) / Ed(0+)), we
can express Q as follows:
ð
SA
Eo (z)|ax dz
Q~
V
ð6Þ
ax SA
|½Ed (0z )|(1{R(0z )){Enet (H)
~ |
V
at
In Eq. 6, the irradiance lost to scattering back through the
air–water interface is parameterized through the irradiance
reflectance R, while irradiance lost from the bottom of the
experimental volume is Enet(H). In this general formulation
(Eq. 6), the surface H could represent, for instance, a
physical interface (the bottom of a cell) or an arbitrary
depth within a larger water column. To determine the AQY
for a photochemical reaction in scattering media, we
therefore must know not only Ed (0+,l), ax(l), and at(l; as
in Eq. 3 for non-scattering media; Hu et al. 2002), but also
the irradiance reflectance R(0+,l) and the net irradiance at
the bottom of the experimental volume, Enet(l,H).
Experimental design rationale—To determine Q(l) for
suspensions of absorbing and scattering particles examined
in this study, we express Enet(l,H) in terms of the depthaveraged net irradiance attenuation coefficient ,Knet(l).
(Eq. 7; Fig. 1), which scales approximately directly with the
concentration of absorbing particles and will permit a more
general discussion below. By rewriting Enet(l,H) in terms of
,Knet(l). (Eq. 7), we assume Q(l) is independent of the
optical properties of the experimental volume’s boundaries
(explicit in Eq. 6). This assumption is justified at large
optical depths (i.e., deep enough that nearly all incident
light is attenuated within the sample volume, as in our
experiments), where the optical properties of the bottom
surface at H have only a small effect on Q(l),
Q(l)~
ax (l) SA
|
ðEd (l,0z )|(1{R(l,0z ))
at (l)
V
|(1{e{vKnet (l)wH )
ð7Þ
Photodissolution rate and quantum yield
1745
Fig. 1. Schematic showing irradiance quantities defined in text. (A) Dashed-line cube shows hypothetical sample volume within
horizontally infinite system. Top face is air–water interface and bottom face is the bottom boundary of the sample volume. Vertical
arrows show upward, downward, and net irradiances just above and below air–water interface and above the bottom boundary.
Wavelength notation is omitted and arrows are vertical for clarity, although in reality irradiances are functions of wavelength and are
integrated over all upward and downward directions. Relative lengths of arrows are proportional to magnitudes of hypothetical
irradiances such that Enet(0+) 5 Enet(02). (B) Plot at right shows definition of mean net irradiance attenuation coefficient ,Knet..
Photochemical reactions under constant irradiance can
be described using either apparent zero- or first-order
kinetics with respect to the reactant (Zepp 1978), depending
on whether the reactant concentration changes appreciably
during irradiation (‘apparent zero-order kinetics’ here is
equivalent to the initial rate approximation for an apparent
first-order reaction). A general rate law (Gordon and Ford
1972), in terms of DOC production from POC, is
d½DOC
~k½POCn
dt
ð8Þ
where n is the reaction order with respect to POC and k is
the rate constant. For an apparent first-order reaction, k is
defined as
lmax
ð
k:½POC|
Q (l)|AQY(l) dl
lmin
lmax
ð
~
ð9Þ
Q(l)|AQY(l) dl
lmin
where k has units of s21 and Q*(l) is the POC-specific light
absorption rate. Our experiments were designed with high
initial POC concentrations, and DOC measurements were
made early in the reaction before major POC loss.
Therefore, Q defined as in Eq. 7 and computed for the
initial suspension (i.e., Q* 3 [POC]initial) can be substituted
in the rate expression and the reaction described with
apparent zero-order kinetics (i.e., with an initial rate
approximation), with k bearing units of mmol DOC
m23 s21. Apparent zero-order rate constants, equivalent
under our approximation to first-order rate constant times
[POC]initial concentration, were determined from the slope
of the regression of DOC concentration against time.
Longer term photodissolution experiments (Kieber et al.
2006; Mayer et al. 2006; Estapa and Mayer 2010) have thus
far been insufficient to determine the apparent reaction
order with respect to POC. In these studies, however, the
observed photodissolution rate, measured both as DOC
production and POC loss, decreased over exposures longer
than 10 h at light intensities similar to this study. This
nonlinearity suggests that a true reaction order is at least
first-order, consistent with depletion of photolabile POC or
another reactant. We justify the initial rate approximation
used in this study by considering the dynamics of sediment
resuspension often encountered in turbid aquatic environments. High SPM concentrations near the surface may only
persist for periods of hours (Grabemann and Krause 1989;
Wright et al. 1990; Kineke et al. 2006). While they persist,
the wind- or current-driven mixed layer likely extends
1746
Estapa et al.
Table 1. Sample summary. SPM 5 suspended particulate matter; wt 5 weight; Fedith 5 combined content of reducible iron oxides
and organic-bound iron; ap 5 mass-specific absorption.
ap (300)
Sample description
Organic carbon content
(6 0.2 mg g21)
Fedith
(6 0.02 wt % Fe)
(6 0.01 m2 g21)
1. Marine SPM, Freshwater Bayou, Apr 2003
2. Marine SPM, Freshwater Bayou, Mar 2008
3. Marine sediment, Atchafalaya River delta, Apr 2008
4. River SPM, Atchafalaya River at Morgan City, Apr 2003
5. River SPM, Mississippi River at Venice, Mar 2004
6. River SPM, Mississippi River at St. Francisville, Dec 2004
20.0
17.4
18.3
17.5
22.3
28.1
nd*
1.45
1.83
2.93
1.87
1.96
0.24
0.22
0.21
0.33
0.23
0.19
* nd 5 insufficient sample quantity for iron oxide determination.
deeper than the depth of light penetration (order 100–
101 cm), providing a continuous resupply of ‘fresh’ POC.
We utilize optical property data and a simple dynamic
model to explore the limits of this initial rate approximation (see Discussion, below).
Experimental—Photodissolution rate measurements
were conducted using suspended particles and sediments
collected from coastal Louisiana and the Mississippi and
Atchafalaya rivers. Sediment grabs (box corer or Ekman
grab) were used to collect unconsolidated surficial sediments (top 0–1 cm), while SPM samples were isolated from
large surface-water volumes via overnight settling or
centrifugation. Particles were freeze-dried and stored dry
and frozen until optical and photochemical analyses, which
should not affect photodissolution extent (Mayer et al.
2006). We selected samples for analysis from a larger set to
represent a wide range of particle mass-specific absorption
coefficients, organic carbon contents, and reducible iron
contents (Table 1). Mass-specific absorption coefficients
were measured in suspension in an integrating sphere
accessory attached to a spectrophotometer, while dithionite-citrate extraction (Loeppert and Inskeep 1996) was
used to determine the combined content of reducible iron
oxides and organic-bound iron (Fedith; see Estapa et al.
2012).
To measure rates of photodissolution, samples were
sieved to , 63 mm, weighed into replicate, precombusted
glass beakers, and covered with precombusted quartz disks.
Reflective foil and matte black tape lined the beakers’ outer
sides and bottoms, respectively. Carbon-clean artificial
seawater (containing Na+, Mg2+, K+, Ca2+, Cl2, and SO2{
4
in seawater proportions to give salinity , 35) was added to
create 3-cm-deep particle suspensions with known mass
concentrations (, 500 mg L21). Suspensions were magnetically stirred from the bottom with carbon-clean, glasscoated stir bars.
Irradiations were conducted in a solar simulator (Atlas
Suntest CPS) equipped with a filtered xenon lamp
replicating the wavelength distribution of solar irradiance.
A thermostatted, recirculating water-bath controlled the
temperature (during most experiments) to 6 1uC (exceptions discussed below). We used 4-h experiments for rate
determination under full-spectrum irradiance in order to
maintain initial conditions. In this case, the concentration
of DOC at time t ([DOC]t) was much less than the initial
concentration of POC ([POC]0), which allowed rate
determination following Eq. 8. Samples of DOC were
collected at 0 h, 0.5 h, 1 h, 2 h, and 4 h after the start of
irradiation. Suspensions were magnetically stirred for
30 min prior to being placed in the solar simulator, so that
initial, non-photochemical, desorption would be assessed in
the t 5 0 h sample. We measured the DOC concentration
by removing sets of duplicate beakers from the solar
simulator, filtering each suspension through a precombusted glass-fiber filter, and analyzing the filtrates in
triplicate with a Shimadzu TOC-VCPH analyzer. POC
concentrations at 0 h and 4 h were determined on the
glass-fiber filters after vapor-phase acidification and
combustion in a Perkin Elmer 2400 elemental analyzer.
We computed the photodissolution rate (mmol DOC
m23 s21) for each experiment as the slope of the best
linear fit to the mean DOC concentration vs. time data.
Uncertainty in the DOC production rate was taken as the
uncertainty in this slope.
We assessed the temperature sensitivity of the photodissolution rate by varying the water-bath temperature in a
series of replicate 4-h irradiations of a single sample
(Atchafalaya River delta sediment, No. 3 in Table 1).
Temperatures within identical, tap-water–filled beakers in
the solar simulator reached steady-state (6 1uC) prior to
substitution by, and irradiation of, the sample beakers
(usually after 30 min with the lamp on and bath
recirculating). Sample beakers removed from the simulator
during each time series were replaced with these tap-water–
filled beakers to maintain the thermal mass inside the
sample chamber. We computed Arrhenius parameters for
the temperature dependence of the reaction by iteratively
fitting the 4-h rate and temperature data to Eq. 10:
ln(k)~1000|
Ea
{ln(A)
Rg T
ð10Þ
where Ea represents the reaction activation energy (kJ mol
DOC21), Rg the gas constant, T the reaction temperature
(u Kelvin), and A the pre-exponential factor (mmol DOC
m23 s21). A Monte Carlo procedure was used to propagate
analytical uncertainty in temperatures and rate constants
into the Arrhenius parameters.
In this study, we measured Ed(0+,l) and the coefficients
of mass-specific absorption (ap (l)) and scattering (bp (l))
of particle samples directly, and used the latter two
parameters to derive ax(l), at(l), and total scattering
coefficients (bt(l)) for experimental suspensions containing
Photodissolution rate and quantum yield
Table 2.
1747
EcolightH input parameters to simulate laboratory conditions. CDOM 5 colored dissolved organic matter.
Parameter
Inherent optical property (IOP) model
IOP model components and concentrations
Mass-specific absorption
Mass-specific beam attenuation
Chlorophyll and CDOM fluorescence and Raman scattering
Wavelength range
Depth range
Bottom depth
Bottom reflectance
Wind speed
Sky model
Sky radiance angular distribution model
Solar zenith
Ozone
Earth–sun distance
Total incident irradiance
Value
Case 2 (routine ‘abcase2.f’)
Seawater (n51.34), mineral particles (500 mg L21), no colored dissolved
organic matter (CDOM) or chlorophyll (concentrations50)
Empirical sample measurements (ap (300): Table 1; ap (l): Estapa
et al. 2012)
Mean field measurements (Eq. 11)
None
300–800 nm, 5-nm increments
0–0.03 m, 0.0004-m increments
0.03 m
0 at all wavelengths (i.e., black bottom)
0 m s21
Semi-empirical, 100% overcast conditions
‘hcnrad’ routine
0u (directly overhead)
Mean annual
Mean annual
Solar simulator output spectrum
only water and known particle mass concentrations. We
computed ax(l) as the product of ap (l) and the particle
mass concentration, at(l) as the sum of ax(l) and aw(l;
known absorption of water), and bt(l) as the product of
bp (l) and the particle mass concentration, plus a small
scattering contribution by water. We then entered Ed (0+,l),
at(l), and bt(l) into a radiative transfer model (described
below) whose outputs included R(0+,l) and ,Knet(l)..
These last two parameters were then used to compute Q(l)
according to Eq. 7.
The solar simulator delivered a constant output irradiance
of 765 W m22 (300–800 nm), approximately equivalent to
subtropical noontime sunlight, during all experiments. A
built-in radiometer calibrated to a National Institute of
Standards and Technology–traceable source monitored and
controlled the lamp power throughout each experiment.
Prior to AQY-determination experiments, we measured the
intensity, wavelength, and spatial distribution of the solar
simulator’s output with nitrate and nitrite actinometers
(Jankowski et al. 1999) and with a spectroradiometer
(Satlantic HyperOCR irradiance sensor). Spectroradiometer
measurements (350–800 nm) were made by collecting timeaveraged measurements of Ed (0+) inside the simulator
sample chamber at varying distances from the four walls.
We prepared actinometer solutions as described by Jankowski et al. (1999), filled beakers identical to those used in
sample irradiations to a depth of 1 cm, and exposed the
beakers to simulated sunlight at 765 W m22 for 1 h. Salicylic
acid production rates, determined via batch fluorescence
measurements, were used to calculate band-integrated
incident irradiances (Jankowski et al. 1999) in the nitrate
(301–328 nm) and nitrite (324–385 nm) absorption bands.
The air–water interfaces of sample suspensions and actinometer solutions, and the sensor face of the radiometer,
were all placed at the same level below the xenon lamp in the
solar simulator. Spatially averaged radiometric and actinometric irradiance measurements agreed with the nominal
irradiance spectrum of the simulator. Irradiances near the
sides of the sample chamber were about 6% below the mean,
while those measured near the center of the sample chamber
were about 6% above the mean.
We used a radiative transfer model (EcolightH, Sequoia
Scientific) to determine depth-averaged Knet(l) and the
irradiance reflectance, R(l) for experimental suspensions.
The spectral absorption and beam attenuation coefficients
of suspensions containing 500 mg L21 suspended sediments
and artificial seawater were used as inputs to the model. We
used the model’s default values for seawater absorption and
scattering (Pope and Fry 1997), and assumed a seawater
index of refraction of 1.34. Mass-specific absorption
spectra of suspended sediment as a function of wavelength
were determined using an integrating sphere accessory
(Estapa et al. 2012). We determined the mass-specific, beam
attenuation spectrum of suspended sediments by regressing
field measurements of spectral beam attenuation against
simultaneously determined SPM concentrations (see below). The angular scattering dependence (volume scattering
function) of suspended particles was approximated with a
Fournier–Forand phase function with a spectrally invariant
particulate backscatter fraction (bb/b) of 0.02—a typical
value for suspended, mostly inorganic particles from this
region (Snyder et al. 2008). In EcolightH, an overcast sky
model approximated the diffusing effect of the reflective
inner walls of the solar simulator. Additional Ecolight
parameters are in Table 2.
Field inherent optical property (IOP) data were collected
in turbid areas of coastal Louisiana in the vicinity of
sediment and SPM sampling locations (Table 1), and they
were used to determine mass-specific beam attenuation and
to validate mass-specific absorption spectra of freeze-dried
samples. In March of 2008, near-surface optical properties
were measured continuously through the flowing seawater
system of the R/V Pelican during a cruise to the turbid
Freshwater Bayou region of coastal Louisiana. Continuously operating instrumentation included a multispectral,
absorption and beam attenuation meter (ac9; WETLabs)
1748
Estapa et al.
cp (l)~cp (531)
Fig. 2. Comparison of mass-specific absorption spectra
measured in the field with an ac9 and on freeze-dried samples in
the lab. Field spectrum (solid line) is the median of measurements
made when SPM concentrations were between 7 mg L21 and
25 mg L21, as described in the text. Dashed lines show mean and
95% confidence interval of 25 freeze-dried samples from the study
region (see Estapa et al. 2012 for details). To account for
scattering-correction uncertainties in the ac9 data, the two spectra
are forced into agreement at 715 nm by adding a constant y-offset
to the field data. Peaks in the field spectrum at 440 nm and 676 nm
are due to phytoplankton pigments not evident in the freeze-dried
samples, which were collected at higher turbidities.
and a three-angle backscatter meter (EcoVSF; WETLabs),
while SPM and colored dissolved organic matter (CDOM)
were sampled by filtration from periodic discrete samples.
Multispectral beam attenuation and absorption data were
corrected for temperature- and salinity-dependent attenuation by water and time binned into 1-min intervals.
Absorption data were additionally corrected for scattering
losses within the ac9 light path according to the third
method of Zaneveld et al. (1994), and CDOM absorption
was subtracted. SPM concentrations were determined
gravimetrically on triplicate, pretared, 0.1-mm pore size,
47-mm-diameter polycarbonate filters (Nucleopore). The
slope of the regression line between the raw backscattering
signal and simultaneously measured SPM concentration
was used to obtain a continuous SPM proxy from the
EcoVSF.
Multispectral, mass-specific particulate beam attenuation (cp (l)) and absorption (ap (l)) were determined from
the slopes of the linear regressions between the SPM
backscatter proxy and corrected beam attenuation and
absorption spectra, over SPM concentrations between
7 mg L21 and 25 mg L21. The lower limit excluded low
suspended-sediment waters, while the upper limit excluded
measurements potentially affected by multiple scattering
within the ac9 light path (Leymarie et al. 2010). A powerlaw function (Eq. 11; Boss et al. 2001) fit the derived cp (l)
vs. wavelength data well, and was used to extrapolate cp
through the 300–800-nm range needed for the radiative
transfer model:
l
531
{0:89
; cp ð531Þ~1 m2 g{1
ð11Þ
The scattering correction applied to the ac9 absorption
data is known to underestimate true absorption in highly
scattering waters and in the presence of significant nonwater absorption at 715 nm (Tzortziou et al. 2006;
Leymarie et al. 2010), while absorption measured with
our integrating sphere setup is subject to minimal scattering
losses (Babin and Stramski 2002). To enable comparison, a
spectrally flat offset was added back into the field data to
force the two data sets into agreement at ap (715). After this
correction, the mean ap (l) spectrum from the field data
was similar in magnitude and variability to the mean massspecific absorption of 25 freeze-dried suspended and
sedimented samples collected from the same region and
measured in the laboratory with the integrating sphere
(Fig. 2; Estapa et al. 2012). Spectral absorption features
characteristic of phytoplankton pigments and degradation
products are present in the field data but not in the freezedried samples, likely because the latter were collected at
higher turbidities.
We computed net irradiance as a function of wavelength
and depth (Enet(l,z)) as the difference between Ed (l,z) and
Eu(l,z) output from the radiative transfer model. The net
irradiance attenuation coefficient Knet(l) was computed as
the mean of the derivative with respect to depth of
ln(Enet(l,z)). We calculated the irradiance reflectance above
the air–water interface (R(l,0+)) as the output ratio of
upward to downward irradiance just above the air–water
interface.
The above-water reflectance R(0+,l) and the mass
(l,
specific, net particulate attenuation coefficient (Kp,net
SPM)) were recomputed in EcolightH for varying concentrations of suspended particles (101–103 mg L21) in order to
assess sensitivity of our results (all from 500 mg L21
suspensions) to varying SPM concentration. Kp,net
(l,SPM)
was computed as
(l,SPM)~ðKnet (l,SPM){aw (l)Þ=SPM
Kp,net
ð12Þ
(Knet(l,SPM) 2 aw(l)) scaled roughly with the SPM
(l,SPM) decreased
concentration (not shown) but Kp,net
slightly as SPM increased (Fig. 3a), reflecting path-length
(l,SPM)
amplification due to increased scattering. Kp,net
was well-approximated by (ap (l) + bbp (l)) / 0.7 (Gordon
1989; Fig. 3a). Above SPM concentrations of about
250 mg L21, R(0+) varied little at ultraviolet (UV) wavelengths (Fig. 3b).
The depth-integrated irradiance absorbed within the
suspension by sediment particles (Q(l)) was finally
computed according to Eq. 7. To determine the spectral
AQY for POC photodissolution, DOC photoproduction
from a single sample (Atchafalaya River delta sediment,
No. 3 in Table 1) was measured under a series of long-pass
wavelength filters (300 nm, 400 nm, 500 nm, 600 nm, and
700 nm) following the method of Johannessen and Miller
(2001). Polyester lighting filters (Lee Filters) shielded the
entire solar-simulator sample chamber. Filter transmission
Photodissolution rate and quantum yield
1749
Fig. 3. R(0+,l) and Kp,net
(l) as a function of SPM concentration. (a) shows Kp,net
(l) for SPM log-spaced at 101, 102, and 103 mg L21
in thin solid lines, with arrow indicating direction of increasing concentration. Thick dashed line shows Kp,net
(l) approximated as
(ap + bbp ) / 0.7 (Gordon 1989). (b) shows R(0+,l) as a function of SPM for selected log-spaced concentrations. Above about 250 mg L21, R(0+)
is relatively invariant at ultraviolet wavelengths.
was constant during the irradiation treatments. In order to
produce measurable DOC concentrations under reduced
photon energies, we increased irradiation times from 4 h
to 5 h, 8 h, 12 h, and 46 h in respective treatments under
400 nm, 500 nm, 600 nm, and 700 nm long-pass filters, but
maintained the same relative sampling frequency and linear
increase of DOC vs. time described above for full spectrum
irradiations. We computed the AQY using a modification
of the statistical curve fit method described by Johannessen
and Miller (2001). Spectral irradiance absorbed in each
treatment was computed from Eq. 7, above, where Ed (l,0+)
was the measured irradiance in the solar simulator, the
relative particulate absorption spectrum ax(l)/at(l) was
computed from the known mass concentration of suspended sediment and the IOPs of sediments and water, and R(l)
and Knet(l) were output from EcolightH. Many AQY
spectra for photoreactions of marine CDOM are welldescribed by an exponential function (Eq. 13, Johannessen
and Miller 2001):
AQY(l)~AQY(290)|e({S(l{290))
ð13Þ
We found that Eq. 13 similarly describes the wavelengthdependence of photodissolution (see Results). We estimated the amplitude (AQY(290)) and slope (S) parameters by
iteratively minimizing Eq. 14 (using MatlabH function
‘fminsearch’):
2
X 1 dDOC
~
~
{AQY(290)exp({S |(l{290)) ð14Þ
y~
Q(l) dt
where Q(l) was computed according to Eq. 7. Finally, a
Monte Carlo procedure was used to propagate the
uncertainty in the DOC production rate into the parameters AQY(290) and S.
To compensate for increasing photodissolution rates as a
function of temperature (Mayer et al. 2006), rates and
AQYs reported below have been recalculated, according
to an Arrhenius rate model (Eq. 15), to a reference
temperature of 25uC (298 K).
Ea
(1=298{1=T)
R
Rate298K ~RateT |e g
{
ð15Þ
where T is the temperature (in u Kelvin) of the suspension,
and Rate298K and RateT refer, respectively, to the rates
at 298K and at the original temperature. The activation
energy (Ea) for photodissolution, determined from rate
measurements between 17u and 29uC, was 32 6 7 kJ mol21
(Fig. 4). The fit of the data to the Arrhenius model was best
above 19uC (Fig. 4). Temperature normalization changed
the rates reported below by , 14% for all samples.
Results
Photodissolution rates during the first hour of irradiation (Table 3; Fig. 5) correlated positively with sample
Fedith content (Table 1; Fig. 5). After the first hour, DOC
photoproduction increased approximately linearly with
time (Fig. 5). The relationship between first-hour photodissolution rate and sample Fedith (n 5 4, R2 5 0.97, p 5
0.013) did not persist for the full 4-h irradiation treatment
(Table 3). Normalization of the 1-h photodissolution rates
to modeled scalar irradiance absorbed by particles at
350 nm (Q(350), mmol DOC nm mE21) did not remove the
dependence on sample Fedith (Fig. 6a). No strong relationships existed between Q(350)-normalized rates over either
time period, and sample organic carbon content or massspecific absorption (Fig. 7b,c).
Irradiance reflectance (R(l)) ranged among samples
from 7–8.5% at 300 nm to 16–18% at 700 nm (Fig. 7a).
The net irradiance attenuation coefficient (Knet(l)) ranged
from 165–222 m21 at 300 nm to 14–19 m21 at 700 nm
(Fig. 7b). Attenuation depths (Fig. 7c) were similar to those
measured by Mayer et al. (2006) via actinometry in 450-mg
1750
Estapa et al.
temperatures ranging from 10u to 35uC (using results of
temperature-sensitivity experiments; see Eq. 15 and Fig. 4).
The subsequently computed AQY spectral slope, S, did not
vary as a function of temperature, but the pre-exponential
term AQY(290) did. We therefore recomputed AQY(290)
separately for all broadband experiments (for which
accurate temperatures were available), while assuming a
constant value of S determined from long-pass cut-off
experiments (Eq. 16).
d½DOC
dt
AQY(290)~ {S(l{290)
e
|Q(l)
Fig. 4. Arrhenius plot of 4-h rates measured in tests of
photodissolution temperature dependence. Arrhenius parameters:
2Ea / Rg (slope) 5 23.9 6 0.9K; ln(A) (intercept) 5 2 6 30 mmol
DOC m23 s21. Error bars are derived from analytical uncertainty
in temperature (6 1uC) and rate.
L21 sediment suspensions, although differences in the solar
simulator and container walls prohibit quantitative comparison. The fraction of surface irradiance reaching the
bottom of the suspensions (Fig. 7d) shows that . 90% of ,
400-nm light passing the air–water interface was absorbed
within the suspension, and . 75% at wavelengths , 500 nm,
thus confirming ‘optically thick’ conditions at photochemically important UV and blue wavelengths.
In suspensions exposed to long-pass–filtered sunlight,
photodissolution rates decreased as the wavelength of
incident irradiance increased (Table 4), but visible wavelengths drove a substantial portion of the reaction. POC
photodissolved at a rate of 2.4 6 0.2 mmol DOC m23 s21
under the 300-nm long-pass filter, and 1.2 6 0.1 mmol DOC
m23 s21 under the 400-nm filter. Smaller, but still
significant, reaction rates were observed under 500 nm,
600 nm, and 700 nm long-pass filters (Table 4).
Space limitations did not permit us to measure temperature in wavelength-filtered experiments to the same
precision achieved in the broadband irradiations. We tested
the sensitivity of AQY(290) and S to uncertainty in the
temperature measurement by scaling long-pass–filtered
photodissolution rates (measured at 19 6 2uC) to
ð16Þ
The AQY spectral slope determined from long-pass–
filtered experiments was 0.017 6 0.003 nm21, while
AQY(290) values computed from broadband irradiations
of samples listed in Table 3 ranged from 0.0007 mmol DOC
mE21 to 0.0009 mmol DOC mE21 (Table 3). The resulting
AQY spectra for all samples are shown in Fig. 8. As a test,
the derived AQY spectrum for the Atchafalaya delta
sediment (Sample 3 in Table 1) was used to predict DOC
production rates in the individual long-pass filter experiments listed in Table 4 (which were conducted with the
same sediment sample). The fit between measured and
modeled DOC production rates in these experiments had a
root mean-squared error of 0.094 mmol DOC m23 s21,
which was less than the analytical uncertainty in most of
the rate determination experiments (Table 3). The alternate, ‘quasi-exponential’ AQY functional form of Bélanger
et al. (2006) was tested, but the minimization function did
not converge (regardless of the initial solution guess used
for the iterative fitting routine), suggesting that the quasiexponential model does not describe these data well.
Discussion
From a mechanistic standpoint, the photodissolution
AQY reported here cannot be easily compared to published
dissolved-phase marine AQY spectra, because the relative
contents of absorbing-but-photoresistant compounds in
POC are probably different from those in CDOM (Zafiriou
2002). From a modeling standpoint, however, such
comparisons are useful, because AQY spectra allow us to
estimate in situ photoreaction rates from a waterbody’s
radiometric and optical properties (Bélanger et al. 2008).
This rationale led us to compute the photodissolution AQY
Table 3. Photodissolution rates (95% CI) adjusted to T 5 25uC (mmol DOC m23 s21). SPM 5 suspended particulate matter; AQY 5
apparent quantum yield.
Sample description
1.
2.
3.
4.
5.
6.
Marine SPM, Freshwater Bayou, Apr 2003
Marine SPM, Freshwater Bayou, Mar 2008
Marine sediment, Atchafalaya River delta, Apr 2008
River SPM, Atchafalaya River at Morgan City, Apr 2003
River SPM, Mississippi River at Venice, Mar 2005
River SPM, Mississippi River at St. Francisville, Dec 2004
* nd 5 lost 0.5-h sample, so 1-h rate not determined.
1-h rate
4-h rate
Measurement
temperature (uC)
4-h AQY(290)
(31024 mmol mE21)
3.79(0.06)
3.5(0.5)
5.0(0.8)
7.8(0.4)
4.18(0.05)
nd*
2.3(0.5)
2.7(0.1)
3.0(0.3)
2.6(0.7)
3.2(0.2)
2.5(0.4)
25
22
17–31
25
26
25
7(2)
8(1)
9(2)
7(2)
9(2)
7(2)
Photodissolution rate and quantum yield
1751
Fig. 5. Photoproduced DOC concentration, as a function of
time under simulated sunlight, for the four samples in which both
Fedith content and 0–1-h rates were determined (Samples 2–5;
Tables 2 and 3). Solid lines show best fit to [DOC] vs. time data
over 4 h and dashed lines show best fit over the first hour only.
Error bars show analytical uncertainty in DOC concentration.
Root mean-squared (RMS) error for 4-h linear fit ranged from
1 mmol L21 DOC (Sample 2) to 6 mmol L21 DOC (Sample 4).
Samples were mixed for 30 min prior to irradiation; hence, the
non-zero t 5 0 h DOC concentrations.
relative to the total, rather than organic carbon- or Fedithspecific, particulate absorption coefficients.
DIC is the most-abundant, identifiable photoproduct of
CDOM. Figure 9 shows the photodissolution AQY computed in this study alongside AQY spectra for DIC
photoproduction from CDOM, pooled for temperate
‘inshore’ waters (Johannessen and Miller 2001) and Arctic
shelf waters (Bélanger et al. 2006). The AQY for DOC
production via photodissolution is comparable to or
greater than the CDOM AQYs. This similarity suggests
that POC photodissolution may occur in nearshore waters
even in the presence of significant, competing CDOM
Fig. 6. Photodissolution rates normalized to absorbed irradiance at 350 nm (Rate/Q(350)), computed over 1-h (open squares) and
4-h (filled circles). (a) Rate/Q(350) as a function of suspended sediment
Fedith. (b) Rate/Q(350) vs. particulate mass-specific absorption at
350 nm. (c) Rate/Q(350) as a function of suspended sediment organic
carbon fraction. Error bars represent analytical uncertainties and
labels correspond to sample numbers in Tables 1 and 2.
1752
Estapa et al.
Fig. 7. Radiative transfer model outputs. (a) Irradiance reflectance just above the air–water interface (R(0+)), with line shading and
pattern representing samples as organized in Tables 1 and 2. (b) Net irradiance attenuation coefficient (Knet(l)) within 500 mg L21
suspension, with patterns as in (a). (c) Net irradiance within suspension of Atchafalaya River delta sediment (Sample 3) as a function of
depth and selected wavelengths (Enet(z,l)). (d) Fraction of net irradiance (just below surface) reaching the bottom (at 3 cm) of suspension
of Atchafalaya River delta sediment.
absorption. By multiplying the photodissolution AQY by
the mean DIC : DOC photoproduction ratio (0.45) measured in , 20-h irradiations of suspended Louisiana coastal
sediments (Estapa and Mayer 2010), we have computed a
rough AQY estimate for DIC photoproduction from
particles. The magnitude of this estimated AQY is less
certain because it is based on a DIC : DOC photoproduction ratio determined during much longer irradiation
periods (Estapa and Mayer 2010). Even so, this estimate
of particulate DIC photoproduction is comparable to DIC
photoproduction from CDOM in temperate, inshore
waters and Arctic shelf waters (Fig. 9). This calculation
implies that suspended sediment-dominated nearshore
waters may contribute to DIC photoproduction as much
as do CDOM-dominated waters further offshore.
The significant, positive temperature dependence of the
POC photodissolution rate (Fig. 4) suggests that although
the net reaction may be initiated photochemically, it also
involves subsequent, diffusion-limited steps. While direct
photochemical reactions often have very low activation
energies, thermally driven (non-photochemical) reactions
have much higher values. We computed a moderate Ea
value for photodissolution, 32 6 7 kJ mol21, which is
somewhat higher than rarely determined activation energies
of dissolved-phase, marine organic photochemical reactions: 12.2 kJ mol21 for CO photoproduction (Zhang et al.
2006), 13 kJ mol21 for domoic acid photolysis (Bouillon et
al. 2006), and 23 kJ mol21 for near-surface DMS
photolysis (Toole et al. 2003). The phase change inherent
to photodissolution (i.e., desorption or dissolution of
POC) may account for the increased activation energy of
this reaction. The Arrhenius model does not fit the data
well at the low end of the temperature range investigated
(Fig. 4), which may be due to imprecise temperature
control or changing reaction kinetics as a function of
temperature. Our temperature-dependence data agree with
Mayer et al. (2006), who observed that a decrease in
temperature from , 35–40uC to 23uC reduced the 24-h
extent of reaction by one-third. The temperature range
investigated here (17–30uC) includes much of the annual
range of water temperatures observed in the study region
of coastal Louisiana.
We have assumed that the true photodissolution reaction
order is at least first-order with respect to SPM, because
light absorption by particles scales with SPM concentration
and with POC concentration early in the reaction (Eq. 9).
We did not explicitly test for a higher order dependence of
photodissolution rate on SPM concentration, but prior
Photodissolution rate and quantum yield
1753
Table 4. Photodissolution rates in long-pass–filtered experiments (T 5 19 6 2uC). DOC 5 dissolved organic carbon.
Cut-off
wavelength (nm)
Length
of experiment (h)
Rate
(mmol-DOC m23 s21)
300
400
500
600
700
4
5
8
12
46
2.4(0.2)
1.2(0.1)
0.26(0.08)
0.14(0.08)
0.08(0.02)
results provide guidance in this respect. Estapa and Mayer
(2010) showed that 24-h photodissolution extents were
identical in continuously stirred beakers and sealed test tubes
shaken by hand every 6–8 h. This equivalence suggests a
minor effect of particle–particle collisions (which depend on
differential particle velocity) on the photodissolution rate.
Unlike CDOM, which photobleaches (absorption decreases) as the light dose increases (Andrews et al. 2000), lengthy
UV exposures negligibly photobleach suspended, primarily
inorganic particles, although dominance of absorption by
iron (Estapa et al. 2012) may mask bleaching of organic
matter. Therefore, consumption of a photoreactive substrate
for this photodissolution reaction might not be accompanied
by measurable photobleaching of the overall particulate
substrate, in contrast to photobleaching of dissolved
materials, which partially explains the dose-dependence of
their photochemical rates (Andrews et al. 2000).
The correlation between Fedith and initial photodissolution rates (Fig. 5) suggests ligand-to-metal charge transfer
reactions in iron–organic complexes, which would photooxidize (and possibly dissolve) the organic ligands while
photoreducing Fe3+ to Fe2+ (Moffett and Zafiriou 1993).
Irradiation of suspended sediments similar to those in this
study caused the total iron content of small (0.7–8.0-mm
diameter) particles to increase, which is consistent with
rapid reoxidation and precipitation of photoreduced iron
(Estapa and Mayer 2010), a process also observed by
Kopáček et al. (2005) during photodegradation of dissolved
organic–metal complexes in lakes. Variable, yet nonnegligible, amounts of organic carbon are adsorbed or
complexed to dithionite-reducible iron in marine sediments
both globally (Lalonde et al. 2012) and in the Mississippi
river and delta system (Mayer et al. unpubl.), consistent
with release of DOC during photoreductive dissolution of
the iron. However, Mayer et al. (2006) found that removal
of dithionite-reducible iron prior to irradiation of suspended sediments did not affect the longer term (tens of hours)
extent of organic carbon photodissolution. From these
findings, we infer that iron photoredox processes may
influence the high photodissolution rates observed in the
first hour of irradiation, but that they become less
important over longer timescales (Table 3; Fig. 6). The
uniformity of Q(350)-normalized 4-h rates reported here
(Fig. 6; mean 6 1 SD 5 0.060 6 0.007 mmol DOC nm
mE21) is consistent with other reports showing remarkably
similar long-term (, 20 h) photodissolution extents (21–
24%) among seasonally and compositionally varying
samples (Mayer et al. 2006).
Fig. 8. Apparent quantum-yield spectra for photodissolution. Thin solid lines show spectra for individual samples listed in
Table 2. Heavy dashed lines denote twice the standard deviation
of the measured sample set. Heavy dotted lines show singlesample analytical uncertainty propagated from uncertainty in
AQY spectral slope parameter (S 5 0.0166 6 0.003 nm21) and
measured photodissolution rates (see Table 2).
Particle suspensions used to determine photodissolution
rates were uniformly mixed through the 3-cm-deep photic
zone (Fig. 7c). To correctly extrapolate these reported rates
to environmental conditions, we must consider local
variations in incident sunlight, mixing depth and intensity,
and suspended particle concentration. Under turbid conditions (. 10 mg L21 suspended sediment), we can
compute attenuation coefficients (Fig. 7b) over varying
suspended sediment concentrations to estimate the depth of
UV- and blue-wavelength light penetration (z5%, here
defined as 3 optical depths, or the 5% light level). A
particle settling from the surface under low-energy conditions, with velocity vs, will spend an amount of time in the
photic zone equivalent to tsettling 5 z5%/vs. Under highenergy conditions, a particle mixed repeatedly from surface
to bottom of the water column or mixed layer (zmixing) will
spend equal amounts of time at all depths , zmixing, and
thus will be exposed to light for the time-integral of the
ratio z5% : zmixing. In Fig. 10 we illustrate these low- and
high-energy cases for suspended sediment concentrations
ranging from 101 mg L21 to 103 mg L21, for zmixing 5 1–
10 m, and integrated over solar irradiance time periods
representative of the summer and winter solstices in coastal
Louisiana. These calculations show that when SPM
concentration exceeds , 102 mg L21, a given particle is
likely exposed to 350-nm light for , 1 h d21 regardless of
water-column depth (for zmixing . 1 m) or day length.
Below this approximate SPM concentration, z5% deepens,
and daily exposure times increasingly depend upon zmixing
and the day length. Therefore, the AQYs computed
here over 4-h time periods likely underestimate rates of
photodissolution occurring in the field, especially at low
1754
Estapa et al.
Fig. 9. Mean 6 twice standard deviation of photodissolution
AQY spectra measured in this study (thick black line and shaded
gray area) along with literature examples of dissolved-phase DIC
production AQY spectra. Long-dashed line shows the mean AQY
for DIC photoproduction from ‘inshore’ CDOM at several pooled
sites (Johannessen and Miller 2001). Dotted lines show AQY
estimates for DIC photoproduction from CDOM at Mackenzie
shelf stations in the Arctic (Bélanger et al. 2006). Thick, gray line
is an estimate of the particulate DIC photoproduction AQY,
based on a 0.45 ratio of DIC : DOC production in Louisiana (LA)
coastal suspended sediments (Estapa and Mayer 2010).
irradiances (e.g., low solar zenith angles and overcast
conditions) and when suspended sediments have high Fedith
content.
The Fedith-dependence of the initial rates measured here
suggests that in situ photodissolution may be enhanced at
sites characterized by episodic resuspension. Aller (2004)
postulated that redox-state oscillation of iron and manganese in muddy sediments, driven by periodic resuspension
events, could enhance benthic organic carbon remineralization during intervening calm periods by renewing
oxidants in the sediments. To the extent that sedimentary
Fe2+ is present in sediments and pore waters, and is
reoxidized during resuspension, oscillating redox conditions might periodically increase reducible iron concentrations at the surface, and thus enhance photodissolution in
the upper water column.
Using an approach combining laboratory experiments
and radiative transfer modeling to characterize light levels
within turbid sediment suspensions, we have presented the
first photodissolution rate measurements under wellcharacterized irradiance conditions. These irradiance- and
absorption-normalized measurements are necessary for
quantitative prediction of the in situ extents of biogeochemically relevant, particle-hosted photoreactions, especially in shallow embayments, estuaries, and marshes where
light absorption rates of suspended particles and dissolved
materials are both significant. Because the AQY for
photodissolution is similar in magnitude to DIC photoproduction from CDOM, relative light absorption rates
of particulate and dissolved organic matter in these
Fig. 10. Schematic showing approximate length of time a hypothetical particle would be exposed to sunlight of 350 nm under
different mixing and optical conditions. Exposure time is assumed to equal the time spent more shallow than three optical depths (i.e., 5%
light level) in a mixed layer or water column 1 m, 5 m, or 10 m thick (thin solid lines; arrow shows direction of increasing depth) with
sediment concentration ranging from 10 mg L21 to 1000 mg L21. (a) assumes 3.4 h d21 of high sun-angle (, 57u) sunlight penetration
into the water, equivalent to winter solstice in coastal Louisiana, and (b) shows 8.7 h d21, equivalent to the summer solstice. Arrows show
direction of increasing depth. Thick lines on both panels show exposure time based only on the canonical settling speed (vs 5 1023 m s21;
Hill et al. [1998]) of an aggregate starting at the surface. Thin dashed lines denote 1-h and 4-h time periods over which reaction rates were
measured in the present study. Optical depths were computed from Knet(l) values for different SPM concentrations but no other
absorbing constituents except pure seawater.
Photodissolution rate and quantum yield
environments should control the relative importance of
photoproduction of DOC from particles and photooxidative DOC loss from the dissolved phase. A rigorous wholewater, photochemical carbon budget will thus require
spatiotemporally resolved estimates of light absorption by
each component.
Acknowledgments
Mary Jane Perry and Mark Wells provided lab space and
equipment for analyses performed in this study. Field data were
collected aboard the R/V Pelican, with ship time and space
provided by Gail Kineke (Office of Naval Research Coastal
Geosciences program, award number N00014-06-1-0718 to R. A.
Dalrymple). We thank two anonymous reviewers and P. J. Hernes
for helpful comments. We would like to acknowledge funding
support from the National Science Foundation Chemical Oceanography program (L.M. and M.L.E.), a National Aeronautics and
Space Administration Earth Systems Science Graduate Fellowship (M.L.E.), and the Office of Naval Research Environmental
Optics program (E.B.).
References
ALLER, R. 2004. Conceptual models of early diagenetic processes:
The muddy seafloor as an unsteady, batch reactor. J. Mar.
Res. 62: 815–835, doi:10.1357/0022240042880837
A NDREWS , S. S., O. C. Z AFIRIOU , AND S. C ARON . 2000.
Photochemical oxygen consumption in marine waters: A
major sink for colored dissolved organic matter? Limnol.
Oceanogr. 45: 267–277, doi:10.4319/lo.2000.45.2.0267
BABIN, M., AND D. STRAMSKI. 2002. Light absorption by aquatic
particles in the near-infrared spectral region. Limnol.
Oceanogr. 47: 911–915, doi:10.4319/lo.2002.47.3.0911
BÉLANGER, S., M. BABIN, AND P. LAROUCHE. 2008. An empirical
ocean color algorithm for estimating the contribution of
chromophoric dissolved organic matter to total light absorption in optically complex waters. J. Geophys. Res. 113:
C04027, doi:10.1029/2007JC004436
———, H. XIE, N. KROTKOV, P. LAROUCHE, W. F. VINCENT, AND
M. BABIN. 2006. Photomineralization of terrigenous dissolved
organic matter in Arctic coastal waters from 1979 to 2003:
Interannual variability and implications of climate change.
Glob. Biogeochem. Cycles 20: GB4005, doi:10.1029/
2006GB002708
BOSS, E., M. S. TWARDOWSKI, AND S. HERRING. 2001. Shape of the
particulate beam attenuation spectrum and its inversion to
obtain the shape of the particulate size distribution. Appl.
Opt. 40: 4885–4893, doi:10.1364/AO.40.004885
BOUILLON, R.-C., T. L. KNIERIM, R. J. KIEBER, S. A. SKRABAL, AND
J. L. C. WRIGHT. 2006. Photodegradation of the algal toxin
domoic acid in natural water matrices. Limnol. Oceanogr. 51:
321–330, doi:10.4319/lo.2006.51.1.0321
ESTAPA, M. L., AND L. M. MAYER. 2010. Photooxidation of
particulate organic matter, carbon/oxygen stoichiometry, and
related photoreactions. Mar. Chem. 122, 138–147, doi:10.1016/
j.marchem.2010.06.003
———, E. BOSS, L. M. MAYER, AND C. S. ROESLER. 2012. Role of
iron and organic carbon in mass-specific light absorption by
particulate matter from Louisiana coastal waters. Limnol.
Oceanogr. 57: 97–112, doi:10.4319/lo.2012.57.1.0097
GERSHUN, A. A. 1939. The light field. J. Math Phys. 18: 51–
151.
GORDON, A. J., AND R. A. FORD. 1972. Chemist’s companion—a
handbook of practical data, techniques, and references. John
Wiley & Sons.
1755
GORDON, H. R. 1989. Can the Lambert–Beer law be applied to
the diffuse attenuation coefficient of ocean water? Limnol.
Oceanogr. 34: 1389–1409, doi:10.4319/lo.1989.34.8.1389
GRABEMANN, L., AND G. KRAUSE. 1989. Transport processes of
suspended matter derived from time series in a tidal estuary. J.
Geophys. Res. 94: 14373–14379, doi:10.1029/JC094iC10p14373
HILL, P. S., J. P. SYVITSKI, E. A. COWAN, AND R. D. POWELL. 1998.
In situ observations of floc settling velocities in Glacier Bay,
Alaska. Mar. Geol. 145: 85–94, doi:10.1016/S0025-3227(97)
00109-6
HU, C., F. E. MULLER-KARGER, AND R. G. ZEPP. 2002. Absorbance, absorption coefficient, and apparent quantum yield:
A comment on common ambiguity in the use of these optical
concepts. Limnol. Oceanogr. 47: 1261–1267.
JANKOWSKI, J. J., D. J. KIEBER, AND K. MOPPER. 1999. Nitrate and
nitrite ultraviolet actinometers. Photochem. Photobiol. 70:
319–328, doi:10.1111/j.1751-1097.1999.tb08143.x
JOHANNESSEN, S. C., AND W. L. MILLER. 2001. Quantum yield for
the photochemical production of dissolved inorganic carbon
in seawater. Mar. Chem. 76: 271–283, doi:10.1016/S0304-4203
(01)00067-6
KIEBER, R. J., R. F. WHITEHEAD, AND S. A. SKRABAL. 2006.
Photochemical production of dissolved organic carbon from
resuspended sediments. Limnol. Oceanogr. 51: 2187–2195,
doi:10.4319/lo.2006.51.5.2187
KINEKE, G. C., E. E. HIGGINS, K. HART, AND D. VELASCO. 2006.
Fine-sediment transport associated with cold-front passages
on the shallow shelf, Gulf of Mexico. Cont. Shelf Res. 26:
2073–2091, doi:10.1016/j.csr.2006.07.023
K OPÁČEK , J., Š. K LEMENTOVÁ , AND S. A. N ORTON . 2005.
Photochemical production of ionic and particulate aluminum
and iron in lakes. Environ. Sci. Technol. 39: 3656–3662,
doi:10.1021/es048101a
LALONDE, K., A. MUCCI, A. OUELLET, AND Y. GELINAS. 2012.
Preservation of organic matter in sediments promoted by
iron. Nature 483: 198–200, doi:10.1038/nature10855
LEYMARIE, E., D. DOXARAN, AND M. BABIN. 2010. Uncertainties
associated to measurements of inherent optical properties
in natural waters. Appl. Opt. 49: 5415–5436, doi:10.1364/
AO.49.005415
LOEPPERT, R. H., AND W. P. INSKEEP. 1996. Iron, p. 639–664. In
D. L. Sparks [ed.], Methods of soil analysis. Part 3: Chemical
methods. Soil Science Society of America.
MAYER, L. M., L. L. SCHICK, T. S. BIANCHI, AND L. A. WYSOCKI.
2009a. Photochemical changes in chemical markers of
sedimentary organic matter source and age. Mar. Chem.
113: 123–128, doi:10.1016/j.marchem.2009.01.006
———, ———, K. R. HARDY, AND M. L. ESTAPA. 2009b.
Photodissolution and other photochemical changes upon
irradiation of algal detritus. Limnol. Oceanogr. 54:
1688–1698, doi:10.4319/lo.2009.54.5.1688
———, ———, K. SKORKO, AND E. BOSS. 2006. Photodissolution
of particulate organic matter from sediments. Limnol.
Oceanogr. 51: 1064–1071, doi:10.4319/lo.2006.51.2.1064
———, K. H. THORNTON, AND L. L. SCHICK. 2011. Bioavailability of organic matter photodissolved from coastal
sediments. Aquat. Microb. Ecol. 64: 275–284, doi:10.3354/
ame01530
MOFFETT, J. W., AND O. C. ZAFIRIOU. 1993. The photochemical
decomposition of hydrogen peroxide in surface waters of the
eastern Caribbean and Orinoco River. J. Geophys. Res. 98:
2307–2313, doi:10.1029/92JC02768
POPE, R. M., AND E. S. FRY. 1997. Absorption spectrum (380–
700 nm) of pure water. II. Integrating cavity measurements.
Appl. Opt. 36: 8710–8723, doi:10.1364/AO.36.008710
1756
Estapa et al.
RIGGSBEE, J., C. ORR, D. LEECH, M. DOYLE, AND R. WETZEL. 2008.
Suspended sediments in river ecosystems: Photochemical
sources of dissolved organic carbon, dissolved organic
nitrogen, and adsorptive removal of dissolved iron. J.
Geophys. Res. 113: G03019, doi:10.1029/2007JG000654
RONTANI, J.-F., N. ZABETIA, AND S. G. WAKEHAM. 2011.
Degradation of particulate organic matter in the equatorial
Pacific Ocean: Biotic or abiotic? Limnol. Oceanogr. 56:
333–349, doi:10.4319/lo.2011.56.1.0333
SNYDER, W. A., AND OTHERS. 2008. Optical scattering and
backscattering by organic and inorganic particulates in US
coastal waters. Appl. Opt. 47: 666–677, doi:10.1364/
AO.47.000666
TOOLE, D. A., D. J. KIEBER, R. P. KIENE, D. A. SIEGEL, AND N. B.
NELSON. 2003. Photolysis and the dimethylsulfide (DMS)
summer paradox in the Sargasso Sea. Limnol. Oceanogr. 48:
1088–1100, doi:10.4319/lo.2003.48.3.1088
TZORTZIOU, M., J. R. HERMAN, C. L. GALLEGOS, P. J. NEALE, A.
SUBRAMANIAM, L. W. HARDING, JR., AND Z. AHMAD. 2006.
Bio-optics of the Chesapeake Bay from measurements and
radiative transfer closure. Estuar. Coast. Shelf Sci. 68:
348–362, doi:10.1016/j.ecss.2006.02.016
WAITE, T. 2005. Role of iron in light-induced environmental
processes, p. 255–298. In P. Boule, D. Bahnemann, and P.
Robertson [eds.], Environmental photochemistry part II.
Springer Berlin/Heidelberg.
WRIGHT, L. D., W. J. WISEMAN, JR., Z.-S. YANG, B. D. BORNHOLD,
G. H. KELLER, D. B. PRIOR, AND J. N. SUHAYDA. 1990.
Processes of marine dispersal and deposition of suspended
silts off the modern mouth of the Huanghe (Yellow River).
Cont. Shelf Res. 10: 1–40, doi:10.1016/0278-4343(90)90033-I
XIE, H., AND O. C. ZAFIRIOU. 2009. Evidence for significant
photochemical production of carbon monoxide by particles in
coastal and oligotrophic marine waters. Geophys. Res. Lett.
36: L23606, doi:10.1029/2009GL041158
ZAFIRIOU, O. C. 2002. Sunburnt organic matter: Biogeochemistry
of light-altered substrates. Limnol. Oceanogr. Bull. 11: 69–71.
ZANEVELD, J. R. V., J. C. KITCHEN, AND C. M. MOORE. 1994. The
scattering error correction of reflecting-tube absorption meters.
Proc. SPIE Ocean Opt. XII 2258: 44–55, doi:10.1117/12.190095
ZEPP, R. G. 1978. Quantum yields for reaction of pollutants in
dilute aqueous solution. Water Resour. Res. 12: 327–329.
ZHANG, Y., H. XIE, AND G. CHEN. 2006. Factors affecting the
efficiency of carbon monoxide photoproduction in the St.
Lawrence estuarine system (Canada). Environ. Sci. Technol.
40: 7771–7777, doi:10.1021/es0615268
Associate editor: Peter Hernes
Received: 28 August 2011
Accepted: 14 June 2012
Amended: 27 July 2012